Does Q-Chem support the SDD basis set

Questions
#1

Hi everyone! Does Q-Chem support the SDD basis set, similar to what Gaussian offers? If so, what is its name?

#2

Perhaps not by name. (I think the manual is up-to-date in this respect and also I don’t see it in $QCAUX.) You can use a generalized basis input, perhaps that is what you were trying to do here?

It is pretty trivial for a developer to add new basis sets so you could contact Q-Chem with a feature request if this is important for your work, although that would need to wait for a future release. An alternative is that you can define a user-defined basis set (called USER1 or USER2); see the manual for details.

#3

Thank you John. I will follow your guidance and use a user-defined basis set. However, I believe there is a need for this, as I think SDD is one of the most commonly used basis sets for inorganic complexes in the literature.

#4

I don’t see anything named “SDD” in the basis set exchange; I have a rough idea this is for use with Stuttgart-Dresden ECPs (?) but I don’t know the details. If you can point me to a literature or online reference where the basis set info is provided, I can create a ticket to add this by name…

#5

In this link (https://gaussian.com/basissets/ ), Gaussian explains the SDD basis set. Please search for “SDD” in that link.
The explanation is as follows:

SDD: D95 up to Ar [Dunning77] and Stuttgart/Dresden ECPs on the remainder of the periodic table [Fuentealba82, Szentpaly82, Fuentealba83, Stoll84, Fuentealba85, Wedig86, Dolg87, Igel-Mann88, Dolg89, Schwerdtfeger89, Dolg89a, Andrae90, Dolg91, Kaupp91, Kuechle91, Dolg92, Bergner93, Dolg93, Haeussermann93, Dolg93a, Kuechle94, Nicklass95, Leininger96, Cao01, Cao02]. The SDD, SHF, SDF, MHF, MDF, MWB forms may be used to specify these basis sets/potentials within Gen basis input. Note that the number of core electrons must be specified following the form (e.g., MDF28 for the MDF potential replacing 28 core electrons). OldSDD requests the previous default.
SDDAll: Selects Stuttgart potentials for Z > 2.

or for example in the following paper.

https://doi.org/10.1021/acs.organomet.3c00364
#6

Hi, I think the reason is that the SDD is not a specific type of basis set or ECP, see its convention for every element in “SD ECP Availability” here. Besides, if SDD is used for all atoms in a .gjf file, the Gaussian default setting is 6D 10F (i.e. using Cartesian-type basis functions); but if it is written under Cartesian coordinates, the default setting becomes 5D 7F (i.e. spherical harmonic basis functions). These factors make the SDD definition a little bit of complicated and thus not a built-in option in most of quantum chemistry software packages.

If you want to use Gaussian SDD easily, currently you can use the fch2qchem utility to generate Q-Chem input file from a Gaussian .fch file, which contains basis set (and ECP) data, Cartesian coordinates, keywords and most importantly, the MO coefficients. For example, running the following Shell command

fch2qchem h2o.fch

will generate the file h2o.in and temporary directory h2o in $QCSCRATCH/.

1 Like
#7

A lot of these ECPs (Pseudo | Gaussian.com) are indeed missing by name; I’ve created a ticket but I don’t have an estimate on when it will be resolved. Meanwhile, ECP=GEN will work, you can specify them yourself on a per-element basis.

#8

Thank you very much!